This article is about mobile spherical robots - robotic devices that resemble a ball. If you'd like to find out more about spherical robots - robotic arms with a spherical work envelope, please read my article on spherical robots.
Most mobile robots we have today have wheels. That is an obvious choice as there is considerable amount of knowledge about this type of locomotion. However, more and more possible applications occur where wheeled robots have some flaws.
Mobile spherical robots or robots that, in fact, resemble a ball, could be a solution to some of these problems. There are numerous ways how to design a robot ball. Still, pros and cons are more or less common to all robotic balls.
The above mentioned situations often involve dealing with "difficult" terrain as well. While wheeled robots can cope with it pretty good, the risk of falling over still persists. A mobile spherical robot, on the other hand, can't fall over at all.
Also, it can be quite a task to deploy a wheeled robot without direct human intervention - the landing spot has to be carefully chosen, the robot has to land with wheels down, etc. However a robotic ball can be simply dropped out above the desired position.
Theoretically, all this is great. Nevertheless, there are some downsides as well. Flexibility is one of these issues. If we talk about planetary exploration, the chances are the robot will have to do other tasks alongside observing - taking soil samples, measuring and alike.
While it is relatively easy to make a wheeled platform suitable for such tasks, the chances are it'd be quite troublesome to do so with a robotic ball. Even if it's possible, the robot could loose some of its advantages as a result.
Also, some mobility issues persist. Although robotic balls, when compared to wheeled robots, have shown better performance on such surfaces as snow or sand, they still can't outperform wheeled robots in such locomotion tasks as slope climbing.
There are numerous ways how to design a mobile spherical robot. Most, if not all, approaches to this problem use the sphere's inner mass displacement in one or another way. By doing so, the center of gravity is moved out of the contact area - thus the ball "falls" in that direction.
One way how to achieve this this is to put a hamster in a ball. When the hamster runs (displaces itself in the ball), the ball rolls. The same effect can be achieved if you put in a ball something that drives on its inner surface.
In my opinion, this is the simplest option, it is often used in toys and it is fairly simple to control such robot. However, this may not be the better solution if actual applications are considered - the sphere's body has to be hard. While this is OK with smaller robots, it can get inconvenient if larger robots are considered.
Another solution to this internal mass displacement problem is a pendulum. If a pendulum is attached to a fixed axis then it is possible to create the robot's body using many different approaches, as nothing has to rest on the inner surface of the sphere.
So, when the pendulum is tilted the robot moves forward/backward or leans to the side. A derivation of this approach is when a non-fixed axis is used. In this way additional mobility and turning capabilities are ensured.
There are some other ways how to move the internal mass. In either case, these approaches, when used in a "raw" manner, doesn't provide good ability to overcome obstacles or slopes. However, the solutions can be found as demonstrated by this prototype:
I'll tell a word or two about this robot a bit later. Apart from the above mentioned approaches, there are other, more exotic approaches, such as deformable bodies, use of natural forces, or the combination of these two and the above mentioned.
In either case, there's still plenty of room for improvement and innovation in the design of mobile spherical robots.
In my opinion, every mobile spherical robot is a notable development - if you make one, congratulations, you help to develop the humankind. Still, I'd like to point out some "robot balls" that I view more or less as milestones of their kind.
The tumbleweed ball is a mobile spherical robot prototype developed by scientists from NASA's Jet Propulsion Lab. The story goes like this - the scientists experimented in the desert with a rover that had large, inflated spherical wheels.
One of these wheels went rogue and demonstrated remarkable desert traversing capability as it rolled away blown by the wind. Scientists who observed this quickly realized the potential of a large, inflated spherical vehicle.
Since then, numerous prototype vehicles have been made using different materials and approaches to the steering problem. The idea is pretty ingenious and I really believe that devices like the tumbleweed ball will play a significant role in planetary exploration. You can read more on this robot in this article at NASA.
This robot has the same roots as the previous one - space exploration. Initially, this was a prototype robot developed in the framework of a project funded by ESA - European Space Agency. However, now it's designed and sold commercially as a surveillance solution.
As I understand, this robot is a pendulum driven mobile spherical robot with a fixed axis. The axis is a bit extended to provide place for cameras and other possible hardware such as sensors, microphones or speakers. Also, it should be noted that it can travel on various surfaces such as sand, snow, mud and even liquids.
The most notable thing about this particular spherical robot is that it is the first commercially available utility robot of this type. Correct me if I'm wrong. Of course, there are toy robots of this type and there was the now-discontinued entertainment robot Q-Taro from SONY, but Groundbot is the most serious one in my opinion.
If you search for other youtube videos from this user ("RotundusRobotics"), you'll see that it can float on water, as well as run on snow.
This is the robot that climbs a step in the video under the "Design" section of this article. It is designed by Greg Schroll, who was a student at MIT at the time. The robot and its inventor received the Next Generation Award at the 2009 Popular Mechanics Breakthrough Awards.
So, what's so different about this robot that makes it so unique? As I mentioned above, most spherical robot designs have some serious flaws - they can't overcome obstacles, negotiate steep inclinations and such.
This prototype, on the other hand, is capable of such behavior by using gyros. As far as I know, it can only overcome one obstacle at a time. However, the concept works and chances are it can be improved further. Who knows, maybe we'll soon see a stair-climbing mobile spherical robot.